reddyn12
reddyn12
CUDA=1 python3 examples/mamba.py With async - out of memory error within the load dict call. Also, weird your getting False for output matched on CLANG. The contiguous fixed that issue....
I'm dumb. Didn't see the prompt arg, ignore the output match part. If async works on tiny15, could it be a method that works for multi gpu setups? Cuz I...
I have 3080. I'll be home tmr and can check again then. I'm p sure I had cuda 12.4 when I nvidia-smi.
https://github.com/reddyn12/tinygrad/tree/mamba_new is the fresh branch @nimlgen. Still getting the same error: ``` reddyn@Nikhil-3080:/mnt/d/Code/tinygrad$ CUDA=1 python3 examples/mamba.py Special tokens have been added in the vocabulary, make sure the associated word embeddings...
same error with ``` CUDA=1 python3 examples/gpt2.py ```
mamba and gpt2 work when I use school's compute cluster. Have a strong feeling its single gpu system related
Btw, I had to comment out the gym import cuase there was a problem with that too.
This branch timed: ```PYTHONPATH=. python3 test/test_pattern_matcher.py 0.84s user 0.31s system 251% cpu 0.459 total``` https://github.com/reddyn12/tinygrad/tree/upat_speed timed: ```PYTHONPATH=. python3 test/test_pattern_matcher.py 0.80s user 0.27s system 187% cpu 0.569 total``` Going to write...
When #4744 gets merged, I'll convert tests to match UPat spec
I realized testing pre UPat and post would be inconvenient if I'm writing the tests in UPat form. That's why I added dict form tests in #4744. Was the closing...